CockroachDB's documentation gives the example
CREATE SCHEDULE core_schedule_label
FOR BACKUP INTO 's3://test/schedule-test-core?AWS_ACCESS_KEY_ID=x&AWS_SECRET_ACCESS_KEY=x'
How can I modify this to use an S3-compatible service like linode rather than AWS?
The format is very similar; you just need to override the endpoint with your actual linode endpoint. A linode s3 URI can look like
CREATE SCHEDULE my_own_backup_schedule FOR BACKUP INTO 's3://test/schedule-test-core?AWS_ACCESS_KEY_ID=accesskeyid&AWS_SECRET_ACCESS_KEY=secret&AWS_REGION=us-east-1&AWS_ENDPOINT=https://us-east-1.linodeobjects.com'
Note that the AWS_ENDPOINT is just the host, not the full endpoint with the bucket name. On older versions of CockroachDB, providing the bucket name in AWS_ENDPOINT (like AWS_ENDPOINT=https://us-east-1.linodeobjects.com/test/schedule-test-core) worked, but in 22.1+ backups created like that may get the error "failed to list s3 bucket". You can fix this issue by creating a new backup schedule formatted as above and adding WITH SCHEDULE OPTIONS ignore_existing_backups so that you don't get an error like unexpected error occurred when checking for existing backups in s3 from validations in current code trying to use the older URI.